Skip to content

Model & Compute Platform — bounded contexts

Generated from the architecture model

This page is rendered from model/**/*.yaml by tools/docs_model during mkdocs build. Edit the YAML, not this page; hand-written design lives on the linked service, platform and plane pages.

flowchart LR
    subgraph model_compute["Model & Compute Platform"]
        BC_07["BC-07<br/>Model Access"]
        BC_41["BC-41<br/>Compute Pools & Placement"]
    end
Hold "Alt" / "Option" to enable pan & zoom
Context Authoritative state Services Status
BC-07 Model Access Model profiles, route plans, invocation receipts 6 target
BC-41 Compute Pools & Placement Model-serving compute pools, placement decisions, warm state 1 target

BC-07 Model Access

Authoritative state: Model profiles, route plans, invocation receipts | Service | Granularity | Host | Status | |---|---|---|---| | AzureOpenAIIntegrationService | external / canonical | — | unmodelled | | ModelGatewayHost | standalone host | — | target | | ModelPolicyService | module of host | ModelGatewayHost | unmodelled | | ModelRouterService | module of host | ModelGatewayHost | target | | OllamaIntegrationService | external / canonical | — | unmodelled | | OpenAIIntegrationService | external / canonical | — | unmodelled |

Events produced: ModelProfileActivated, ModelProfileRetired, DeploymentStateChanged, CapabilityEvidenceRecorded, RoutePlanResolved, ModelInvoked, ModelFallbackUsed

BC-41 Compute Pools & Placement

Authoritative state: Model-serving compute pools, placement decisions, warm state | Service | Granularity | Host | Status | |---|---|---|---| | ComputePoolService | module of host | ModelGatewayHost | target |

Events produced: ComputePoolRegistered, ComputePoolStateChanged, ModelWarmed, ModelEvicted, PlacementDecided